CIC-3 Home | Visitor
Info | Teams | Projects
| Publications | Internal
Navigation
Map | Calendar | News
| Phone Book | Search
COMPRESSION ERROR
Peak Signal to Noise Ratios (in dB), Event tatra249a
----------------------------------------------------
Target Entropy (bits)
|
|
original
|
4.0
|
3.0
|
2.0
|
1.5
|
1.0
|
0.75
|
0.50
|
0.25
|
| Instantaneous Power
(primary peak)
|
58.25
|
57.14
|
55.68
|
61.57
|
75.55
|
79.42
|
113.38
|
153.06
|
236.78
|
| Power Autocorrelation
(secondary peak)
|
38.29
|
36.93
|
35.45
|
42.57
|
54.95
|
50.86
|
51.56
|
43.44
|
44.20
|
(Peak SNR for Instantaneous Power = 20 log Contrast)
Note how the SNR rises as the bitrate decreases; this is due to the fact
that signal compression tends to remove low amplitude and high-frequency
signal content, which has the effect of removing background noise in favor
of higher-energy signal components. This type of noise suppression
is a common phenomenon in many lossy data compression applications.
(We conjecture that the decline in SNR for the secondary autocorrelation
peak at rates below 0.75 bit is attributable to overall losses in signal
energy.)
For this reason we won't present error plots for power contrast or autocorrelation
secondary peak SNR. Moreover, we'll only present error plots
for the autocorrelation secondary peak lag for signals in the "TIPP" (i.e.,
double impulse) category.
Detaild Procedure
L O S A L A M O S
N A T I O N A L L A B O R A T O R Y
Operated by the University
of California for the Department of Energy
CIC - 3
| c3webmeister@c3.lanl.gov | LANL
| Copyright © UC
1998 | Disclaimer
| 25 June 1998
Shane Crockett <m991368@nadn.navy.mil>
and C M Brislawn <brislawn@jacobs.c3.lanl.gov>